On January 12, 2026 NIST's Center for AI Standards and Innovation (CAISI) issued a Request for Information on security considerations for AI agents — systems that plan and take autonomous actions affecting real-world systems — with comments due March 9, 2026 (docket NIST-2025-0035). It asked about adversarial threats such as indirect prompt injection, data poisoning and insecure deployment, misaligned behavior such as specification gaming, measurement approaches for agent security, and safeguards for constraining and monitoring agent access in production.
| Document | CAISI RFI on AI agent security (2026) — Request for Information: Security Considerations for Artificial Intelligence Agents |
| Issued by | National Institute of Standards and Technology — AI Risk Management Framework |
| Type | Consultation |
| Status | Proposed · comment period closed |
| Published | Jan 12, 2026 |
| Comment deadline | Mar 9, 2026 |
| Applies to | Public RFI; input sought from AI developers, deployers and security researchers. Informs forthcoming NIST/CAISI agent-security guidance relevant to any bank deploying agentic AI. |
| Official source | federalregister.gov ↗ |
| Use cases | Generative & agentic AI · Cybersecurity · AI governance (general) |
What are the key points of CAISI RFI on AI agent security (2026)?
- Issued January 12, 2026; Federal Register notice 2026-00206; comments closed March 9, 2026.
- Defines the scope as AI agents that combine model outputs with software functionality to plan and act autonomously.
- Topic areas: adversarial threats (indirect prompt injection), model vulnerabilities (poisoning, insecure deployment), misaligned behavior (specification gaming, unintended harmful actions), measurement and evaluation, and deployment safeguards (constraining and monitoring access).
- Feeds CAISI's AI Agent Standards Initiative launched February 17, 2026, which also includes an NCCoE concept paper on software and AI agent identity and authorization and sector listening sessions naming financial services.
- No guidance document had been published from the RFI as of August 2026.
What did CAISI RFI on AI agent security (2026) change for banks?
The RFI is the first formal US federal fact-finding specifically on securing AI agents and signals what NIST guidance will cover: agent identity and authorization, least-privilege tool access, monitoring and kill-switches, and prompt-injection defenses. Banks piloting agents for operations, servicing or coding can use its topic list as a preview of the control expectations examiners are likely to import.
What did NIST's 2026 AI agent security RFI ask about?
Threats to AI agents (indirect prompt injection, data poisoning, insecure deployment), misaligned agent behavior, ways to measure agent security, and safeguards for constraining and monitoring agents in production. Comments closed March 9, 2026.
What is CAISI?
The Center for AI Standards and Innovation within NIST, renamed from the US AI Safety Institute in 2025. It leads NIST's AI agent standards work, including the January 2026 security RFI and the February 2026 AI Agent Standards Initiative.
| Date | Document | Status |
|---|---|---|
| Apr 7, 2026 | AI RMF critical-infrastructure profile (concept note) — Concept Note: AI RMF Profile on Trustworthy AI in Critical Infrastructure | Proposed |
| Dec 16, 2025 | NIST IR 8596 (Cyber AI Profile) — Cybersecurity Framework Profile for Artificial Intelligence (Cyber AI Profile), NIST IR 8596 — preliminary draft | Proposed |
| Aug 14, 2025 | NIST COSAiS control overlays — Control Overlays for Securing AI Systems (COSAiS): SP 800-53 overlays for generative, predictive and agentic AI | Proposed |
| Mar 24, 2025 | NIST AI 100-2e2025 (Adversarial ML) — Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations (NIST AI 100-2e2025) | Final |
| Jul 26, 2024 | NIST AI 600-1 (Generative AI Profile) — Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile (NIST AI 600-1) | In force |
| Jan 26, 2023 | NIST AI RMF 1.0 — Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST AI 100-1 | In force |
Follow every document these regulators publish
6 curated AI stories for banking executives · Every morning · Free
Subscribe to BankingNewsAI →